基于DSP的无位置传感器BLDCM的模糊控制系统研究

上海交通大学硕士学位论文基于DSP的无位置传感器BLDCM的模糊控制系统研究姓名：蔡磊申请学位级别：硕士专业：电机与电气指导教师：谢宝昌200507012005732005732005731DSPBLDCM2PIDMATLAB/SIMULINKDSP3STUDYONFUZZYCONTROLSYSTEMOFSENSORLESSBLDCMBASEDONDSPABSTRACTIncomparisonwithtraditionalDCmotor,BLDCM(BrushlessDirectCurrentMotor)hasabetterreliability,smallersize,higherefficiency,andcanworkwellunderharshenvironments.Inrecentyears,withthedevelopmentofmocro-processorsandelectrical-electronictechnics,BLDCMhasawideapplicationinindustry,agriculture,traffic,aerospace,andnationaldefence,andit’sofsignificantvaluetodoresearchonBLDCM.ResearchonsensorlessBLDCMcontrolsystemsmainlyfocusedonthewaytogettherotorpositionsignalwithoutsensors,andthecontrolstrategyusedinspeed¤tregulationloops.TraditionalsensorlessmethodsalwayscheckthezerocrossingoftheBEMFofnon-fedphasetogettherotorpositionsignal.Butwhenatlow4speed,theBEMFwillberarelydetected.Thesensorlessmethodpresentedinthispapercanhelpthesystemachieveabetterperformanceatlowspeed,whichwasverifiedbytheexperimentalresults.Themostimportantpartofthispaperistosetupfuzzycontroller.ComparedwithtraditionalPIDcontroller,fuzzycontrollerhasmoreparameterssuchastheshapesofmembershipfunction,thenumberoffuzzyvariablesoffuzzysubsetsofinputsandoutput,andgainsofuniversetransformation.Toselecttheparametersofthefuzzycontroller,asimulationmodelofBLDCMfuzzycontrolsystemwasbuiltusingMATLAB/SIMULINK.Toreducethestaticerrorunderfuzzycontrol,sometechnicalmeasuresweretakenandprovedtobeeffectiveintheexperimentalresult.Intheprocessofthecontrolsystem’sbuilding,softwaresolutionsweretakeninsomecasestoreplacehardwarecomponentssoastosimplifythecircuitandtoreducethetotalcostofthesystem.Inthesectionofhardwaredesigning,theoveralldesignofhardwarewasgivenatthebeginningandthendiscussedindetailtherealizationofsomefunctionalcircuits.WedesignedtwoPCBboards,oneiscontrolboardandtheotherisdrivingboard.Theseparatingofcontrolcircuitandpowercircuitmadethehardwaremorereliableandsimplified.Inthesectionofsoftware5organization,thealgorithmofspeedfuzzycontrolandinterruptionsubroutineswereemphasizedon.TheexperimentalresultsshowedthemoreexcellentperformanceoffuzzycontrolBLDCMsystemthanthatofaPIDone.ItcanbeallegedthatfuzzycontrolBLDCMsystemwillhaveamorebrightfuture.KEYWORDS:BLDCM,sensorlesspositioning,fuzzycontrol,open-loopstartingstrategy,variableregion81.1BrushlessDirectCurrentMotorBLDCM1917Boliger1955D.HarrisonNdFeBIPMDSP9PID10BLDCMpscontroldriveM1M3M5M4M6M21-1Fig1-1BLDCMsystem1.21-1PS1-21-2B1-2M1M6ABM1ABM6ABFa11C120o60oA60oBCM6M2M1ACM2AC60o120060o120o90o30o60o120o120oAASNSNBCXYZBCXYZ1-2Fig1-2CommutationofBLDCMddFaFa1260o1-3M3M2M3M4M5M4M5M6M1M6M1M21-11-1Table1-1Two-phaseconductingsequencesofWye-connected3-phaseBLDCM1.3ABCBCABM2M1M3M4M5M613dqabc12345+++=nnncbacbaccbcabcbbaacabacbacbacbaUUUeeeiiidtdLMMMLMMMLiiiRRRUUU0000001.11.1aUbUcUaRbRcRaLbLcLabMbaMbcMcbMcaMacMaibiciaebecenURa=Rb=Rc=RMMMMMMMLLLLaccacbbcbaabcba=========1.114()++−+=nnncbacbacbacbaUUUeeeiiidtdMLiiiRUUU1.20=++cbaiii1.31.21.33eee3UUUUcbacban++−++=1.4ccbbaaeieieieP++=1.5ωω)(ccbbaacieieieT++=1.6JBTTdtdLe)(ωω−−=1.7ωθpdtd=1.81.7LTBJ1.8θp1120oπ2πSa(θ)S(θ)Sb(θ)Sc(θ)0θ1-3Fig.1-3Shapefunctionwaveforms1-115120o1-3=)()()(θθθωcbaecbaSSSKeee1.916BLDCM2.12.1.1MicroLinearML4425/26ML44282-130o30o2-1Fig.2-1Sampleofterminalvoltage1730o2-1U1R1R2C1D1U2A/DPWMf212112122RRRRfCjRUU++=π2.1U2U1211212RRCRfRarctg+=πα2.2θ∆=30oαα30oθ∆θ∆60o+θ∆2.1.21.218()()∑+=∑−=∑=∞=∞=∞=...5,3,10...5,3,10...5,3,1120sin120sinsinnwnecnwnebnwneankkenkkenkkeθωθωθω2.3kwnn2.3∑+=++=...5,3333sin33sin3nneecbankkkkeeeθωθω2.432.1.3BLDCM900195%BLDCM()()[]∑∞=−−=11212sinkkkaSθθ2.52.5()[]()2212126/12sin24−−=−kkakππ2.6)(θaS)(θbS)(θcS[]−=−=−=++=)()()()()()()()()()()()()(000000310θθθθθθθθθθθθθSSSSSSSSSSSSSccbbaacba2.72.7)(0θS1/32-2Fig.2-2Shapefunctionsaftertransformationπ2πSa0(θ)S(θ)Sb0(θ)Sc0(θ)0θS0(θ)-4/34/320)(0θaS)(0θbS)(0θcS4/32-22.5~2.7331)(scbanEuuuu−++=2.82.8ωθ⋅=)(03SKEes2.81.6()+−+=000cbacbacbacbaeeeiiidtdMLiiiRuuu2.92.9[ea0eb0ec0]tEs3)(0θaS)(0θbS)(0θcS2-212.5~2.721)(0θS2.92-32-4ADADG1(s)=Vo(s)/Vi(s)1)(/)(23221222132121++++=sCRCRCRsCCRRRRsG2.10ADG1(s)2.92.10R1R2R3C1C2ViVo2-4Fig.2-4Two-orderRCfilter2-3Fig.2-3Principlediagramofspeedobserver22()[]−+−=)()()()()()()()()()(1000sIsIsIsGsMLRsVsVsVsEsEsEcbacoboaocba2.112.11G1(s)Iabc(s)[R+L-Ms]121/333004232.2e1.9()()()⋅⋅=θθθωcrbrarecbafffkeee2.12keω()θabcrfθ()θfωωθω[2]ω241.2+−+=)()()()(θλθλθλcrbrarcbacbacbadtdiiidtdMLiiiRuuu2.131.8⋅⋅+−+=)()()()(θθθθωcrbrarecbacbacbafffddkiiidtdMLiiiRuuu2.14)(),(),(θθθcrbrarfff120ouaubuc2.14ab}{θθθωdffdkdtiidMLiiRUbrarebabaab)()()()()(−⋅⋅+−−+−=2.15{}θθθωdffdkdtiidMLiiRUcrbrecbcbbc)()()()()(−⋅⋅+−−+−=2.16{}θθθωdffdkdtiidMLiiRUarcreacacab)()()()()(−⋅⋅+−−+−=2.17abH)(θ(){}θθθθdffdHbrarab)()